www.nature.com/scientificreports OPEN Microbiota and metabolome responses in the cecum and serum of broiler chickens fed with plant essential oils or virginiamycin Yan Chen, Jun Wang, Longfei Yu, Tianyue Xu & Nianhua Zhu* This study investigated the cecal microbiota and serum metabolite profle of chickens fed with plant essential oils (PEO) or virginiamycin (VIRG) using high-throughput 16S rRNA gene sequencing and untargeted metabolomics approach. The main aim of this work was to explore the biochemical mechanisms involved in the improved growth performance of antibiotics and their alternatives in animal production. The results showed that both PEO and VIRG treatment signifcantly increased the relative abundance of phyla Bacteroidetes and decreased the abundance of phyla Firmicutes and genus of Lactobacillus in cecal microbiota of chickens. Compared to the control group (CT group), the relative abundance of genus of Alistipes, unclassifed Rikenellaceae, Roseburia, and Anaeroplasma was enriched in the PEO group; that of genus Bacteroides, Lachnospiraceae, and unclassifed Enterobacteriaceae was enriched in the cecal microbiota of the VIRG group. Untargeted metabolomics analyses revealed that the PEO treatment modifed 102 metabolites and 3 KEGG pathways (primary bile acid biosynthesis and phenylalanine metabolism) in the cecal microbiota, and 81 metabolites and relevant KEGG pathways (fructose and mannose metabolism, biosynthesis of unsaturated fatty acids, and linoleic acid.) in the serum of the chicken. Compared to the CT group, VIRG treatment group difered 217 metabolites and 10 KEGG pathways in cecal contents and 142 metabolites and 7 KEGG pathways in serum of chickens. Pearson’s correlation analysis showed that phyla Bacteroidetes and genus of Bacteroides, Alistipes, and unclassifed Rikenellaceae (in the VIRG and PE group) were positively correlated with many lipid metabolites. However, phyla Firmicutes and genera Lactobacillus (higher in the CT group) were negatively correlated with the lipid and thymine metabolism, and positively correlated with hydroxyisocaproic acid, cytosine, and taurine. This study shows that dietary supplementation with PEO and VIRG altered the composition and metabolism profle of the cecal microbiota, modifed the serum metabolism profle. Te intestinal microbiota, the population of microorganisms that inhabit the intestine, plays an important role in the intestinal morphology, immunity, nutrient digestion and absorption, and host health1–3. Many studies have demonstrated that intestinal microbiota participates in many metabolic pathways, such as lipid metabolism and amino acid synthesis4,5. Te mechanism by which PEO promote growth of may be alter gut microfora and hence improved absorption of nutrients6, increase absorption of micronutrients in the small intestine7 and reducing the deleterious efects of the microbial metabolites8–10. Tere are many alternatives to antibiotics, such as acidifers, probiotics, oligosaccharides, and plant extracts, which play a growth promoting role by regulating gut microbes in pig and poultry11. Plant essential oils (PEO), which can be extracted from plants by steam distillation, extrusion, or solvent extraction12,13, serve as alternatives for antibiotics used in feed for their safety and limited residual efects5,13–15. Many studies have shown that PEO can decrease the number of E. coli and increase the amount of Lactobacilli in the intestine of broilers5,16–18. Tis has benefcial efects on the intestinal morphology and barrier, and the antiox- idant in animals19. Also, Altop et al. (2018) reported the interactions of host gut–microbiota and co-metabolism Key Laboratory of Animal Nutrition, College of animal science and technology, Jiangxi Agricultural University, Nanchang, Jiangxi, 330045, P. R. of China. *email: [email protected] SCIENTIFIC REPORTS | (2020) 10:5382 | https://doi.org/10.1038/s41598-020-60135-x 1 www.nature.com/scientificreports/ www.nature.com/scientificreports on chickens that were fed with essential oils20. However, only a few studies have investigated the efects of PEO or antibiotics on the cecum microbiota and serum metabolites profle of the chicken. In this study, the cecal microbial composition and metabolites in the cecum and serum of chickens that were fed with virginiaymicin (VIRG) or PEO were investigated through 16S rRNA gene sequencing and untargeted mass spectrometry (UPLC-Q-TOF/MS) metabolomic analysis. Comparing the diferences in the cecal microbial composition and the potential chemical metabolites might help us understand how the dietary antibiotics or PEO alter the intestinal metabolism and health of the broiler chickens. Results Diversity and structure of the cecal microbiota. High-throughput sequencing obtained 176,636 qual- ity-controlled reads from 18 cecum samples from three treatments (PEO group,VIRG group and CT group). Afer denoising, removing chimeras, and fltering low quality sequences, the sample had an average of 44,120 sequences, the average length of each sequence being 443 nucleotides. Based on the 97% identity level, these sequences were decomposed into 1,971 operational taxonomic units (OTUs). Te 16S rRNA gene amplicon sequencing results were deposited in the Sequence Read Archive of the NCBI (accession number PRJNA553851). Te diversity of the cecal microbiota in the three groups are shown in Fig. 1. Compared to the CT and the PEO group, the Chao1, ACE, and Shannon diversity indices were observed to be signifcantly lower in the VIRG group. No diferences in the diversity indices (Shannon, Chao1, and Observed species) were observed between the PEO and the CT group (Fig. 1A and Supplementary Table 1A). Principal Component Analysis based on EUCLIDEAN distance showed that the cecal microbes, among the three treatments of CT, PEO, and VIRG groups, formed distinct clusters. Te treatment groups were well separated with 48.96% and 27.11% variation by the principal components PC1 and PC2, respectively (Fig. 1B). PLS-DA model shows signifcant separation and discrimina- tion, indicating that the PLS-DA model can be used to identify diferences between groups. Dietary treatment with PEO or VIRG (Fig. 2 and Supplementary Table 1B) resulted in a change in the cecal microbial composition. At the phylum level (Fig. 2A), Firmicutes was the most predominant phylum (more than 60%), followed by Bacteroidetes (10–20%), Actinobacteria (2–6%), and Verrucomicrobia (1%). Additionally, there were more than 6% unclassifed bacteria in the cecal microbiota of chickens. Compared to the CT group, the relative abundance of phyla Bacteroidetes increased while that of phyla Firmicutes decreased in the cecal microbi- ota of the PEO and the VIRG groups. At the genus level (Fig. 2B), Lactobacillus, Faecalibacterium, Bacteroidaceae, and unclassifed Rikenellaceae were the predominant genus in the cecal microbiota of the chicken. Te relative abundance of Lactobacillus was lower, and that of unclassifed Rikenellaceae was higher in the cecal microtia of the PEO group. Te genus Bacteroides was higher in the VIRG group than in the CT group. Te LEfSE test results (Fig. 2, LDA > 2, P < 0.05) showed that the proportion of genera Alistipes, unclassifed Rikenellaceae, Anaeroplasma, unclassifed Bacillaceae, and Roseburia were higher in the cecal microbiota of the PE group (Fig. 2C). Te relative abundance of phyla Bacteroidetes and the genus of Bacteroides, unclassifed Enterobacteriaceae, unclassifed Lachnospiraceae, and Anaerofustis were higher in cecal microbiota of VIRG group. Te proportion of phyla Firmicutes and Actionbacteria and genus of Lactobacillus, streptococcus, unclas- sifed Erysipelotrichaceae, coriobacteriaceae, and streptococcaceae were higher in cecal microbiota of CT group (Fig. 2C). Efects of dietary plan essential oils or virginiamycin on caecum and serum metabolites. A non-targeted LC–MS-based metabolomics platform was used to analyze the cecum contents and the serum metabolite profles of chicken fed supplemented with PEO or VIRG. According to the variable importance in the projection (VIP) value > 1, in 95% jack-knifed confdence intervals and P < 0.05, detailed information about the diferent biomarker metabolites has been shown in Supplementary Table 2. Compared to the CT group, 102 diferent metabolites with 53 LC-MS/MS(ESI+) and 49 LC-MS/MS(ESI-), were detected in cecum contents of the PEO group; 81 diferent metabolites (30 ESI+ and 51 ESI-) were detected in the serum of the PEO group. VIRG group have 217 diferent metabolites (96 ESI+ and 121 ESI-) in cecum contents, and 142 diferent metabolites (82 ESI+ and 60 ESI-) in the serum as compared to the CT group. Tere were 305 diferent metabolites (148 ESI+ and 157 ESI-) in cecum contents, and 119 diferent metabolites (58 ESI+ and 61 ESI-) in the serum between PEO and VIRG groups. Next, these diferent metabolites were conclusively identifed using the in-house databases. Figure 3 shows the diferent metabolites in the cecal contents and serum. Compared to the CT group, dietary supplementation with PEO resulted in six more lipid metabolites and two more Carbohydrate metabolites in the cecal contents. Tese included 1-Oleoyl-sn-glycero-3-phosphocholine (LysoPC(18:1(9Z))), 1-Palmitoyl-2-hydroxy-sn-glycero- 3-phosphoethanolamine (LysoPE(16:0/0:0)), 1-Palmitoyl-sn-glycero-3-phosphocholine, Glycocholic acid, Caprylic acid, and Chenodeoxycholate. Amino acid related metabolites (Isoleucyl-valine, Hydroxyproline, beta-Homoproline) decreased in the cecal contents of chicken fed with PEO (Fig. 3A). Additionally,